Capek Characteristic Modes

Characteristic Modes
Part I: Introduction
Miloslav Čapek
Department of Electromagnetic Field

Czech Technical University in Prague, Czech Republic
miloslav.capek@fel.cvut.cz
Seminar
May 2, 2018
Čapek, M. Characteristic Modes – Part I: Introduction 1 / 39

Conventionally, characteristic modes In are
defined as
XIn = λn RIn ,
in which Z = R + jX is the impedance matrix.
Dominant characteristic mode of helicopter

model discretized into 18989 basis functions,
ka = 1/2, decomposed into characteristic
modes in AToM in 47 s.

defined as
XIn = λn RIn ,
However, who knows:

I What is impedance matrix and how to get it?
I What the hell are the characteristic modes? Dominant characteristic mode of helicopter
I Why are they of our interest? ka = 1/2, decomposed into characteristic

defined as
XIn = λn RIn ,
However, who knows:

Therefore, . . .
. . . the characteristic mode theory is to be systematically derived.

defined as
XIn = λn RIn ,
However, who knows:

Therefore, . . .
. . . the characteristic mode theory is to be systematically derived.
Disclaimer: There will be equations! Brace yourself and be prepared. . .

Outline
1 Necessary Background
2 Discretization and Method of Moments
3 Definition of Characteristic Modes
4 Properties of Characteristic Modes
5 Activities at the Department J 1 (r, t)
6 Concluding Remarks
J 2 (r, t)
This talk concerns:
I electric currents in vacuum (generalization is, however,
straightforward),
I time-harmonic quantities, i.e., A (r, t) = Re {A (r) exp (jωt)}.

Necessary Background
Electric Field Integral Equation1
σ→∞
(PEC)
Original problem.
1
R. F. Harrington, Time-Harmonic Electromagnetic Fields, 2nd ed. Wiley – IEEE Press, 2001
k E s (r)
σ→∞
(PEC)
k E i (r)
Original problem.

n̂ × E s r 0 + E i r 0 = 0, r0 ∈ Ω
1
k E s (r)
σ→∞ 0 , µ0
(PEC)
Ω Ω
k E i (r)
Original problem. Equivalent problem.

n̂ × E s r 0 + E i r 0 = 0, r0 ∈ Ω
1
k E s (r) k E s (r)
σ→∞ 0 , µ0
(PEC)
Ω Ω J (r 0 )
k E i (r)
Original problem. Equivalent problem.

n̂ × E s r 0 + E i r 0 = 0, r0 ∈ Ω −n̂ × n̂ × E i r 0 = Z (J ) , J = J r0
1
Electric Field Integral Equation – Problem Formalization
Key role of the impedance operator Z (J )

n̂ × n̂ × E s r 0 = Z (J ) = −n̂ × n̂ × (jωA + ∇ϕ) .
Substituting for Lorenz gauge-calibrated potentials2 A and ϕ gives

Z

Z (J ) = jkZ0 G r, r 0 · J r 0 dS
Ω
2
J. D. Jackson, Classical Electrodynamics, 3rd ed. Wiley, 1998
Electric Field Integral Equation – Problem Formalization
Key role of the impedance operator Z (J )

n̂ × n̂ × E s r 0 = Z (J ) = −n̂ × n̂ × (jωA + ∇ϕ) .
Substituting for Lorenz gauge-calibrated potentials2 A and ϕ gives

Z Z
0
0
1 e−jk|r0 −r|
Z (J ) = jkZ0 G r, r · J r dS = jkZ0 1 + 2 ∇∇ · J r 0 dS,
k 4π |r 0 − r|
Ω Ω
I Impedance operator Z is linear, symmetric (reciprocal, thus non-Hermitian).

I Alternative formulation MFIE3 , common extension towards CFIE3 .
2
3
W. C. Gibson, The Method of Moments in Electromagnetics, 2nd ed. Chapman and Hall/CRC, 2014
Discretization and Method of Moments
Dicretization of the Problem
Only canonical bodies can typically be evaluated analytically.

Problem −n̂ × n̂ × E i (r 0 ) = Z (J ) has to be solved numerically!
0 , µ0
Equivalent problem.
4
J. A. De Loera, J. Rambau, and F. Santos, Triangulations – Structures for Algorithms and Applications. Berlin,
Germany: Springer, 2010
Dicretization of the Problem
Only canonical bodies can typically be evaluated analytically.

Problem −n̂ × n̂ × E i (r 0 ) = Z (J ) has to be solved numerically!
I Discretization4 Ω → ΩT is needed (nontrivial task!)
0 , µ0
ΩT
Equivalent problem. Triangularized domain ΩT .
4
J. A. De Loera, J. Rambau, and F. Santos, Triangulations – Structures for Algorithms and Applications. Berlin,
Germany: Springer, 2010
Representation of the Operator (Recap.)
Engineers like linear systems
L (f ) = h.
f L h
I Typically unsolvable for f in the present state Linear system with input f and
output h.
(how to invert L?).

Engineers like linear systems
L (f ) = h.
f L h
I Typically unsolvable for f in the present state Linear system with input f and
output h.
(how to invert L?).
Representation in a basis {ψ n } and linearity of operator L readily gives5

N
X
In L (ψ n ) = h.
n=1
I One equation for N unknowns → still unsolvable.
5
R. F. Harrington, Field Computation by Moment Methods. Wiley – IEEE Press, 1993
Using proper inner product h·, ·i and N tests from left, we get
N
X
In hχn , L (ψ n )i = hχn , hi,
n=1
i.e., in matrix form the method of moments5 relation reads
LI = H.
5
Algebraic Solution – Method of Moments
Piecewise basis functions6

ln ± Tn+ ln ρ− P−
ψ n (r) = ρ (r) ρ+
n n
2A±n n
P+ A−
n A+
n
n
Tn−
z r
x O y
RWG basis function ψ n .
6
S. M. Rao, D. R. Wilton, and A. W. Glisson, “Electromagnetic scattering by surfaces of arbitrary shape”, IEEE Trans.
Antennas Propag., vol. 30, no. 3, pp. 409–418, 1982. doi: 10.1109/TAP.1982.1142818

ψ n (r) = ρ (r) ρ+
n n
2A±n n
P+ A−
n A+
n
n
Tn−
are applied to approximate J (r) as r
z
N
X
J (r) ≈ In ψ n (r) . x O y
n=1
6

ψ n (r) = ρ (r) ρ+
n n
2A±n n
P+ A−
n A+
n
n
Tn−
are applied to approximate J (r) as r
z
N
X
J (r) ≈ In ψ n (r) . x O y
n=1
Galerkin testing7 , i.e., {χn } = {ψ n }, is performed as

Z
ψ · Z (ψ) dS = hψ, Z (ψ)i ≡ Z = [Zpq ] ∈ CN ×N .
Ω
6
7
P. M. Morse and H. Feshbach, Methods of Theoretical Physics. McGraw-Hill, 1953
From Impedance Operator Z to Impedance Matrix Z
The impedance matrix Z reads

Z Z Z

Zpq = ψ p · Z ψ q dS = jkZ0 ψ p (r 1 ) · G (r 1 , r 2 ) · ψ q (r 2 ) dS1 dS2 .
Ω Ω Ω


Z Z Z

Ω Ω Ω
I We say8 : “Matrix Z is the impedance operator Z represented in {ψ n } basis.”
8
C. Cohen-Tannoudji, B. Diu, and F. Laloe, Quantum Mechanics (2 vol. set), 1st ed. Wiley, 1992

Z Z Z

Ω Ω Ω

I Matrix Z can be calculated, e.g., in AToM9 (plenty of numerical techniques and
tricks should/can be used10 ).
8
9
(2017). Antenna Toolbox for MATLAB (AToM), Czech Technical University in Prague, [Online]. Available:
www.antennatoolbox.com
10
A. F. Peterson, S. L. Ray, and R. Mittra, Computational Methods for Electromagnetics. Wiley – IEEE Press, 1998

Z Z Z

Ω Ω Ω

I Generally11 , impedance matrix Z inherits properties of impedance operator Z.
• Symmetric, complex-valued.
8
9
10
11
N. I. Akhiezer and I. M. Glazman, Theory of Linear Operators in Hilbert Space, 2nd ed. Dover, 1993

Z Z Z

Ω Ω Ω

I Generally11 , impedance matrix Z inherits properties of impedance operator Z.
• Symmetric, complex-valued.
I Matrix Z completely describe the scattering properties of radiator ΩT .
8
9
10
11
Two Hilbert Space Representations12
analytics numerics
govering expression −n̂ × n̂ × E i = Z (J ) V = ZI
type of solution exact approximate
quantities operators, functions matrices, vectors
usage limited general
solution of the system no inverse∗ I = Z−1 V
representation Z = hr, Z (r)i Z = hψ, Z (ψ)i
bilinear form (for Z) p = hJ , Z (J )i p ≈ IH ZI
Z
∗
hf , gi = f ∗ (x) · g (x) dx, AH = AT
Ω
12
Example: Complex Power Balance
Quantity being important in the following.

I Continuous form13 using operator Z

Z Z
1 ∗ 1
− J · E s dS = J ∗ · Z (J ) dS = Prad + 2jω (Wm − We ) .
2 2
Ω Ω
13
I Continuous form13 using operator Z

Z Z
1 ∗ 1
− J · E s dS = J ∗ · Z (J ) dS = Prad + 2jω (Wm − We ) .
2 2
Ω Ω
I Algebraic form14 using matrix Z

Z
1 1
J ∗ · Z (J ) dS ≈ IH ZI = Prad + 2jω (Wm − We ) .
2 2
Ω
13
14
Definition of Characteristic Modes
Diagonalization of the Impedance Operator/Matrix
Motivation
Describe behavior of a scatterer Ω without feeding considered.

Diagonalization of the Impedance Operator/Matrix
Motivation
Describe behavior of a scatterer Ω without feeding considered.
Diagonalization of the impedance matrix Z:

   
Z11 Z12 Z13 ··· Z1N ν1 0 0 ··· 0
   
 Z21 Z22 Z23 ··· Z2N   0 ν2 0 ··· 0 
   
   
 Z31 Z32 Z33 ··· Z3N   0 0 ν3 ··· 0 
   
   
 .. .. .. .. ..   .. .. .. .. .. 
 . . . . .   . . . . . 
   
ZN 1 ZN 2 ZN 3 · · · ZN N 0 0 0 ··· νN
Impedance matrix Z. Yet-unknown diagonalization of impedance matrix Z.

Impedance Operator Represented in Spherical Harmonics
z Example: Let us represent the impedance

Z
operator
in a basis15 of spherical harmonics J sh
n (ϑ, ϕ, a) ∈R
ϑ r D E Z
sh sh
Jm, Z Jn = J sh
m · Z J sh
n dS,
0 Ω
ϕ
y
x
Spherical shell of radius a.
15
This can be understood as solving method of moments analytically with spherical harmonics as basis functions.
16
J. A. Stratton, Electromagnetic Theory. Wiley – IEEE Press, 2007
Impedance Operator Represented in Spherical Harmonics
z Example: Let us represent the impedance

Z
operator
in a basis15 of spherical harmonics J sh
n (ϑ, ϕ, a) ∈R
ϑ r D E Z
sh sh
Jm, Z Jn = J sh
m · Z J sh
n dS,
0 Ω
ϕ
y
x This representation gives diagonal matrix16 , i.e.,
D E
J sh
m, Z Jn
sh sh
= 2 Prad,n sh
+ 2jω Wm,n − We,nsh
δmn ,
δmn = 1 ⇔ m = n ∧ δmn = 0 ⇔ m 6= n.
Spherical shell of radius a.
15
This can be understood as solving method of moments analytically with spherical harmonics as basis functions.
16
J. A. Stratton, Electromagnetic Theory. Wiley – IEEE Press, 2007
Characteristic Modes of Spherical Shell

Formula for spherical shell normalized to unitary radiated power (no units)

sh sh − W sh
!
J m , Z J sh
n 2ω Wm,n e,n

sh = 1 + j δmn ,
J m , R J sh
n
sh
Prad,n
where Z = R + jX



sh sh − W sh
!
J m , Z J sh
n 2ω Wm,n e,n

sh = 1 + j δmn ,
J m , R J sh
n
sh
Prad,n
where Z = R + jX or, alternatively (without problems due to division by zero),

D E D E
J sh sh
m, R Jn + jX J sh
n = 1 + jλsh n J sh
m, R Jn
sh
δmn .



sh sh − W sh
!
J m , Z J sh
n 2ω Wm,n e,n

sh = 1 + j δmn ,
J m , R J sh
n
sh
Prad,n

D E D E
J sh sh
m, R Jn + jX J sh
m, R Jn
sh
δmn .
Linearity of the impedance operator allows to write

D E D E
sh sh sh sh sh
Jm, X Jn = λn J m , R J n δmn ,



sh sh − W sh
!
J m , Z J sh
n 2ω Wm,n e,n

sh = 1 + j δmn ,
J m , R J sh
n
sh
Prad,n

D E D E
J sh sh
m, R Jn + jX J sh
m, R Jn
sh
δmn .
Linearity of the impedance operator allows to write

D E D E
sh sh sh sh sh
Jm, X Jn = λn J m , R J n δmn ,
which is solved for all n ∈ {1, . . . , ∞} via generalized eigenvalue problem (GEP)

X J sh n = λ sh
n R J sh
n .

Characteristic Modes – Generalization
Characteristic modes for arbitrarily shaped body are defined as GEP
X (J n ) = λn R (J n ).

X (J n ) = λn R (J n ).
Algebraic form17 is commonly used instead
XIn = λn RIn ,
with Z = R + jX being the impedance matrix and In ∈ RN ×1 being expansion

coefficients.
17
Only six canonical bodies can, in principle, be solved analytically.
X (J n ) = λn R (J n ).
Algebraic form17 is commonly used instead
XIn = λn RIn ,
with Z = R + jX being the impedance matrix and In ∈ RN ×1 being expansion

coefficients.
We know that GEP18 is capable to diagonalize both R and X operators19 .
I Behavior solely described by the impedance operator/matrix.
I No feeding present (neither E i , nor V)!
17
Only six canonical bodies can, in principle, be solved analytically.
18
G. H. Golub and C. F. Van Loan, Matrix Computations. Johns Hopkins University Press, 2012
19
Generally, only two operators can simultaneously be diagonalized. Separable bodies are exceptional!
Historical Overview23
1948 First mention of diagonalization of the scattering operator by

Montgomery et al.20
23
The literature will be closely reviewed later.
20
C. G. Montgomery, R. H. Dicke, and E. M. Purcell, Principles of Microwave Circuits. New York, United States:
McGraw-Hill, 1948

Montgomery et al.20
1968 Rigorously introduced by Garbacz21 as field/current solutions E n /J n

with orthogonal far-fields and radiating unitary power.
23
20
McGraw-Hill, 1948
21
R. J. Garbacz, “A generalized expansion for radiated and scattered fields”, PhD thesis, Department of Electrical
Engineering, Ohio State University, 1968

Montgomery et al.20

1971 Generalized by Harrington and Mautz22 for antenna problem using

impedance matrix Z as ZIn = νn MIn , νn ≡ 1 + jλn , M ≡ R.
23
20
McGraw-Hill, 1948
21
22
R. F. Harrington and J. R. Mautz, “Theory of characteristic modes for conducting bodies”, IEEE Trans. Antennas
Propag., vol. 19, no. 5, pp. 622–628, 1971. doi: 10.1109/TAP.1971.1139999

Montgomery et al.20
I proposal
I analytical form
1971 Generalized by Harrington and Mautz22 for antenna problem using
impedance matrix Z as ZIn = νn MIn , νn ≡ 1 + jλn , M ≡ R.
I algebraic form
23
20
McGraw-Hill, 1948
21
22
Properties of Characteristic Modes
Characteristic Numbers λn
Rayleigh quotient24 is defined as

Z
J ∗n · X (J n ) dS
2ω (Wm,n − We,n ) IH XIn
λn = Ω
Z = ≈ nH .
Prad,n In RIn
J ∗n · R (J n ) dS
Ω
24
J. H. Wilkinson, The Algebraic Eigenvalue Problem. Oxford University Press, 1988
Characteristic Numbers λn
Rayleigh quotient24 is defined as

Z
J ∗n · X (J n ) dS
2ω (Wm,n − We,n ) IH XIn
λn = Ω
Z = ≈ nH .
Prad,n In RIn
J ∗n · R (J n ) dS
Ω
I Notice Prad,n > 0 ⇒ R 0 is required.

I Eigenvalues λn represent the stationary points25 .
24
J. H. Wilkinson, The Algebraic Eigenvalue Problem. Oxford University Press, 1988
25
J. Nocedal and S. Wright, Numerical Optimization. New York, United States: Springer, 2006
Physical Meaning of Characteristic Numbers λn
Characteristic modes can be classified as

Wm,n > We,n ⇒ λn > 0 mode is of inductive nature,
Wm,n < We,n ⇒ λn < 0 mode is of capacitive nature,
Wm,n = We,n ⇒ λn = 0 mode is in resonance26 .
I To get current to the resonance, let us combine modes with λn < 0 and λn > 0.
26
Resonance of a modal current impressed in vacuum is doubtful. It cannot be excited independently.
Physical Meaning of Characteristic Numbers λn
Characteristic modes can be classified as

Wm,n > We,n ⇒ λn > 0 mode is of inductive nature,
Wm,n < We,n ⇒ λn < 0 mode is of capacitive nature,
Wm,n = We,n ⇒ λn = 0 mode is in resonance26 .
I To get current to the resonance, let us combine modes with λn < 0 and λn > 0.
Knowledge in group theory27 gives understanding of

I degeneracies,
I crossings,
I crossing avoidances.
26
Resonance of a modal current impressed in vacuum is doubtful. It cannot be excited independently.
27
R. McWeeny, Symmetry, An Introduction to Group Theory and Its Applications. Dover, 2002, isbn: 0-486-42182-1
Cardinality of a Set of Characteristic Modes
For a radiator Ω 3 r 0 of finite extent:
Jn λn
Mapping between current densities and their complex power ratios.

Jn λn
Characteristic modes can freely be combined as

X
J= αn J n .
n

J 2ω(Wm −We )
Jn λn
Prad
Characteristic modes can freely be combined as

X
J= αn J n .
n
I Set of characteristic modes is infinite, but countable.

I Set of all currents has higher cardinality (uncountable).
Characteristic Numbers λn for Spherical Shell
20
15
10
5
λn
0
Wm < We
−5
TM modes
−10
−15
−20
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5
ka
Characteristic Numbers λn for Spherical Shell
20
15 TE modes
10
5
Wm > We
λn
0
Wm < We
−5
TM modes
−10
−15
−20
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5
ka
Characteristic Eigenangles δn
Characteristic angles28 δn scale the dynamics of λn ∈ (−∞, ∞)

◦ 1
δn = 180 1 − arctan (λn ) .
π
to δn ∈ (90◦ , 270◦ ).
28
E. Newman, “Small antenna location synthesis using characteristic modes”, IEEE Trans. Antennas Propag., vol. 27,
no. 4, pp. 530–531, 1979. doi: 10.1109/TAP.1979.1142116
Characteristic Eigenangles δn
Characteristic angles28 δn scale the dynamics of λn ∈ (−∞, ∞)

◦ 1
δn = 180 1 − arctan (λn ) .
π
to δn ∈ (90◦ , 270◦ ).
Similarly as for characteristic numbers:

λn > 0 ⇒ δn < 180◦ mode is of inductive nature,
λn < 0 ⇒ δn > 180◦ mode is of capacitive nature,
λn = 0 ⇒ δn = 180◦ mode is in resonance.
28
E. Newman, “Small antenna location synthesis using characteristic modes”, IEEE Trans. Antennas Propag., vol. 27,
no. 4, pp. 530–531, 1979. doi: 10.1109/TAP.1979.1142116
Characteristic Eigenangles δn for Spherical Shell
260
TM modes
240
220
200
Wm < We
δn
180
160
140
120
100
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5
ka

Characteristic Eigenangles δn for Spherical Shell
260
TM modes
240
220
200
Wm < We
δn
180
Wm > We
160
TE modes
140
120
100
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5
ka

Spherical Shell Solved Numerically
While simplest canonical body, spherical shell has plenty of potential issues, e.g.,
I degenerate eigenspace29 , D (l) = 2l + 1,

I conformity of spherical surface with
commonly used basis functions,
I internal resonances30 (λn → ∞),
I computationally demanding evaluation31 .
Spherical shell made of 194 triangles.
29
K. R. Schab and J. T. Bernhard, “A group theory rule for predicting eigenvalue crossings in characteristic mode
analyses”, IEEE Antennas Wireless Propag. Lett., no. 16, pp. 944–947, 2017, K. R. Schab, J. M. Outwater Jr.,
M. W. Young, et al., “Eigenvalue crossing avoidance in characteristic modes”, IEEE Trans. Antennas Propag., vol. 64, no. 7,
pp. 2617–2627, 2016. doi: 10.1109/TAP.2016.2550098
30
T. K. Sarkar, E. Mokole, and M. Salazar-Palma, “An expose on internal resonance, external resonance and
characteristic modes”, IEEE Trans. Antennas Propag., vol. 64, no. 11, pp. 4695–4702, 2016. doi: 10.1109/TAP.2016.2598281
31
Having no junctions, spherical shell (and other closed objects) has highest possible ratio between number of basis
functions (unknowns) and triangles (3/2).
Characteristic Modes J n of Spherical Shell
Dominant capacitive characteristic mode (spherical Dominant inductive characteristic mode (spherical
harmonic TM10 ). harmonic TE10 ).

Properties of Characteristic Quantities
Orthogonality relations32
1 H
I ZIn = (1 + jλn ) δmn ,
2 m
r Z2π Zπ
1 0
F ∗m · F n sin ϑ dϑ dϕ = δmn ,
2 µ0
0 0
i.e., orthogonalization of modal complex power and modal far-fields.
32
R. F. Harrington and J. R. Mautz, “Computation of characteristic modes for conducting bodies”, IEEE Trans.
Summation of Characteristic Modes
Summation formula X
J= αn J n
n

Summation formula X
J= αn J n
n
derived using linearity of the impedance operator and orthogonality of

characteristic modes as
X hJ n , E i i X Vi
n
J= Jn = Jn
n
hJ n , Z (J n )i n
1 + jλn

Summation formula X
J= αn J n
n
derived using linearity of the impedance operator and orthogonality of

characteristic modes as
X hJ n , E i i X Vi
n
J= Jn = Jn
n
hJ n , Z (J n )i n
1 + jλn
with Vni being modal excitation coefficient33 and Mn = 1/ |1 + jλn | being modal
significance coefficient34 .
I Connection between “external” and “modal” worlds.
33
34
M. Cabedo-Fabres, E. Antonino-Daviu, A. Valero-Nogueira, et al., “The theory of characteristic modes revisited: A
contribution to the design of antennas for modern applications”, IEEE Antennas Propag. Mag., vol. 49, no. 5, pp. 52–68,
2007. doi: 10.1109/MAP.2007.4395295
Bonus: On the Inversion of Z operator
Summation formula slightly rearranged (Dirac notation used, i.e., L|f i = |gi)
X hJ n |E i i
|J i = |J n i
n
hJ n |Z|J n i

X hJ n |E i i
|J i = |J n i
n
hJ n |Z|J n i
Let’s do some magic. . .

X |J n ihJ n |
|J i = | − n̂ × n̂ × E i i
n
hJ n |Z|J n i
and compare with the defining formula | − n̂ × n̂ × E i i = Z|J i.

X hJ n |E i i
|J i = |J n i
n
hJ n |Z|J n i
Let’s do some magic. . .

X |J n ihJ n |
|J i = | − n̂ × n̂ × E i i
n
hJ n |Z|J n i
and compare with the defining formula | − n̂ × n̂ × E i i = Z|J i.

We get
X |J n ihJ n |
Z −1 ≡ .
n
hJ n |Z|J n i
I But, do not even try to calculate!

Good theory needs time. . .
26
25
Number of publications
21
20
15 13
10 10
10 9
8
5
5 4
2 2
1 1 1 1 1 1 1 1
0
1970 1975 1980 1985 1990 1995 2000 2005 2010 2015
Year
Number of publications in IEEE Trans. Antennas and Propagation and IEEE Antennas and Wireless
Propagation Letters. WOS database queries: “characteristic modes”, “theory of characteristic modes”, and
“characteristic mode analysis”, relevance checked manually.

Notes on Characteristic Modes
1. Characteristic modes derived for apertures35 , dielectrics36 , magnetics36 , etc.
35
R. F. Harrington and J. R. Mautz, “Characteristic modes for aperture problems”, IEEE Trans. Microw. Theory Techn.,
vol. 33, no. 6, pp. 500–505, 1985. doi: 10.1109/TMTT.1985.1133105
36
R. F. Harrington, J. R. Mautz, and Y. Chang, “Characteristic modes for dielectric and magnetic bodies”, IEEE Trans.

2. Generalization as Inagaki modes37 (orthogonal fields on arbitrary boundary).
35
vol. 33, no. 6, pp. 500–505, 1985. doi: 10.1109/TMTT.1985.1133105
36
37
N. Inagaki and R. J. Garbacz, “Eigenfunctions of composite hermitian operators with application to discrete and
continuous radiating systems”, IEEE Trans. Antennas Propag., vol. 30, no. 4, pp. 571–575, 1982. doi:
10.1109/TAP.1982.1142866

3. Singular expansion method38 , i.e., ZIn = 0.
35
vol. 33, no. 6, pp. 500–505, 1985. doi: 10.1109/TMTT.1985.1133105
36
37
10.1109/TAP.1982.1142866
38
C. Baum, “The singularity expansion method: Background and developments”, IEEE Antennas Propag. Society
Newsletter, vol. 28, no. 4, pp. 14–23, 1986. doi: 10.1109/MAP.1986.27868

b = [hf m , Z (f n )i].
4. Other representations39 applicable, i.e., Z
35
vol. 33, no. 6, pp. 500–505, 1985. doi: 10.1109/TMTT.1985.1133105
36
37
10.1109/TAP.1982.1142866
38
39
A. J. King, “Characteristic mode theory for closely spaced dipole arrays”, PhD thesis, University of Illinois at
Urbana-Champaign, 2015

b = [hf m , Z (f n )i].
5. Other bases40 exist, i.e., AIn = ξn BIn .
35
vol. 33, no. 6, pp. 500–505, 1985. doi: 10.1109/TMTT.1985.1133105
36
37
10.1109/TAP.1982.1142866
38
39
40
L. Jelinek and M. Capek, “Optimal currents on arbitrarily shaped surfaces”, IEEE Trans. Antennas Propag., vol. 65,
no. 1, pp. 329–341, 2017. doi: 10.1109/TAP.2016.2624735

b = [hf m , Z (f n )i].
5. Other bases40 exist, i.e., AIn = ξn BIn .
6. Nowadays implemented in FEKO, CST-MWS, WIPL-D, CEM One, HFSS.
35
vol. 33, no. 6, pp. 500–505, 1985. doi: 10.1109/TMTT.1985.1133105
36
37
10.1109/TAP.1982.1142866
38
39
40
L. Jelinek and M. Capek, “Optimal currents on arbitrarily shaped surfaces”, IEEE Trans. Antennas Propag., vol. 65,
no. 1, pp. 329–341, 2017. doi: 10.1109/TAP.2016.2624735
Activities at the Department
Topics Recently Solved at the Department
I Analytical properties of characteristic modes41 ,

I implementation of characteristic modes42 ,
I benchmarks of commercial and in-house solvers43 ,
I modal Q-factor for antennas44 ,
I radiation efficiency of characteristic modes45 ,
I minimization of Q-factor using characteristic modes46 .
41
M. Capek, P. Hazdra, M. Masek, et al., “Analytical representation of characteristic modes decomposition”, IEEE Trans.
42
M. Capek, P. Hamouz, P. Hazdra, et al., “Implementation of the theory of characteristic modes in Matlab”, IEEE
Antennas Propag. Mag., vol. 55, no. 2, pp. 176–189, 2013. doi: 10.1109/MAP.2013.6529342
43
M. Capek, V. Losenicky, L. Jelinek, et al., “Validating the characteristic modes solvers”, IEEE Trans. Antennas
44
M. Capek, P. Hazdra, and J. Eichler, “A method for the evaluation of radiation Q based on modal approach”, IEEE
Trans. Antennas Propag., vol. 60, no. 10, pp. 4556–4567, 2012. doi: 10.1109/TAP.2012.2207329
45
M. Capek, J. Eichler, and P. Hazdra, “Evaluating radiation efficiency from characteristic currents”, IET Microw.
Antenna P., vol. 9, no. 1, pp. 10–15, 2015. doi: 10.1049/iet-map.2013.0473
46
M. Capek and L. Jelinek, “Optimal composition of modal currents for minimal quality factor Q”, IEEE Trans.
Ongoing Research at the Department
I Improvement of characteristic modes decomposition47 .

I tracking of modal data48 ,
I group theory (symmetries) for tracking and problem reducing49 ,
I characteristic modes for MLFMA50 ,
I interpolation using differentiated GEP,
I characteristic modes for arrays51 .
47
D. Tayli, M. Capek, L. Akrou, et al., “Accurate and efficient evaluation of characteristic modes”,, 2017, submitted,
arXiv:1709.09976. [Online]. Available: https://arxiv.org/abs/1709.09976
48
M. Capek, P. Hazdra, P. Hamouz, et al., “A method for tracking characteristic numbers and vectors”, Prog.
Electromagn. Res. B, vol. 33, pp. 115–134, 2011. doi: 10.2528/PIERB11060209
49
M. Masek, M. Capek, and L. Jelinek, “Modal tracking based on group theory”, in Proceedings of the 12th European
Conference on Antennas and Propagation (EUCAP), 2018, pp. 1–5
50
M. Masek, M. Capek, P. Hazdra, et al., “Characteristic modes of electrically small antennas in the presence of
electrically large platforms”, in Progress In Electromagnetics Research Symposium, St. Petersburg, Russia: IEEE, 2017,
pp. 3733–3738. doi: 10.1109/PIERS.2017.8262407
51
T. Lonsky, P. Hazdra, and J. Kracek, “Characteristic modes of dipole arrays”, IEEE Antennas Wireless Propag. Lett.,
vol. 4, pp. 1–4, 2018. doi: 10.1109/LAWP.2018.2828986
AToM – Antenna Toolbox for MATLAB
AToM developed at the department between 2014 and 2018.

I Capable to calculate matrix Z (and many other matrices),
I capable to calculate the CMs, their tracking, post-processing.
Visit antennatoolbox.com

Special Interested Group

I Established by prof. Lau, Lund University,
I 77 groups worldwide, CTU/Elmag is an active member!
Visit characteristicmodes.org
SIG meeting at EuCAP in London, 2018.

Benchmarking of the CMs Decomposition52
15 exact AToM (1)

FEKO AToM (8)
KS WIPL-D
IDA CEM One
CMC Makarov
10
log10 |λn |
TM modes TE modes
63 48 35 24 15 8 3 3 8 15 24 35 48 63
Visit elmag.org/CMbenchmark
TM/TE mode order
52
M. Capek, V. Losenicky, L. Jelinek, et al., “Validating the characteristic modes solvers”, IEEE Trans. Antennas
Wikipedia Webpage
Visit wikipedia.org/wiki/Characteristic mode analysis

European School of Antennas
Course on
Characteristic Modes: Theory and Applications
Aimed at postgraduate research students and industrial engineers
who want to acquire deep insight into the theory and applications
of characteristic modes.
Visit esoa.webs.upv.es

Concluding Remarks
Summary
Characteristic modes decomposition
XIn = λn RIn
I diagonalizes impedance matrix Z,
I constitutes entire domain basis {In },
I generates orthogonal far-fields {Fn },
I allows compact representation of the radiator ΩT . J1
σ→∞ 0 , µ0
(PEC)
Ω Ω J (r 0 )
ΩT
J2

Questions?
For a complete PDF presentation see capek.elmag.org
Miloslav Čapek
miloslav.capek@fel.cvut.cz
May 3, 2018, v1.10

LATEX, TikZ & PFG used

Capek Characteristic Modes

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Capek Characteristic Modes

Uploaded by

Copyright:

Available Formats

Characteristic Modes

Department of Electromagnetic Field

Čapek, M. Characteristic Modes – Part I: Introduction 1 / 39

Dominant characteristic mode of helicopter

Čapek, M. Characteristic Modes – Part I: Introduction 2 / 39

However, who knows:

Čapek, M. Characteristic Modes – Part I: Introduction 2 / 39

However, who knows:

Čapek, M. Characteristic Modes – Part I: Introduction 2 / 39

However, who knows:

Disclaimer: There will be equations! Brace yourself and be prepared. . .

Čapek, M. Characteristic Modes – Part I: Introduction 3 / 39

Electric Field Integral Equation1

Electric Field Integral Equation1

Electric Field Integral Equation1

Original problem. Equivalent problem.

Electric Field Integral Equation1

Original problem. Equivalent problem.

Electric Field Integral Equation – Problem Formalization

Key role of the impedance operator Z (J )

Substituting for Lorenz gauge-calibrated potentials2 A and ϕ gives

Electric Field Integral Equation – Problem Formalization

Key role of the impedance operator Z (J )

Substituting for Lorenz gauge-calibrated potentials2 A and ϕ gives

I Impedance operator Z is linear, symmetric (reciprocal, thus non-Hermitian).

Dicretization of the Problem

Only canonical bodies can typically be evaluated analytically.

Dicretization of the Problem

Only canonical bodies can typically be evaluated analytically.

Equivalent problem. Triangularized domain ΩT .

Representation of the Operator (Recap.)

Engineers like linear systems

Čapek, M. Characteristic Modes – Part I: Introduction 7 / 39

Representation of the Operator (Recap.)

Engineers like linear systems

Representation in a basis {ψ n } and linearity of operator L readily gives5

I One equation for N unknowns → still unsolvable.

Representation of the Operator (Recap.)

i.e., in matrix form the method of moments5 relation reads

Algebraic Solution – Method of Moments

Piecewise basis functions6

RWG basis function ψ n .

Algebraic Solution – Method of Moments

Piecewise basis functions6

Algebraic Solution – Method of Moments

Piecewise basis functions6

Galerkin testing7 , i.e., {χn } = {ψ n }, is performed as

From Impedance Operator Z to Impedance Matrix Z

The impedance matrix Z reads

Čapek, M. Characteristic Modes – Part I: Introduction 10 / 39

From Impedance Operator Z to Impedance Matrix Z

The impedance matrix Z reads

I We say8 : “Matrix Z is the impedance operator Z represented in {ψ n } basis.”

From Impedance Operator Z to Impedance Matrix Z

The impedance matrix Z reads

I We say8 : “Matrix Z is the impedance operator Z represented in {ψ n } basis.”

From Impedance Operator Z to Impedance Matrix Z

The impedance matrix Z reads

I We say8 : “Matrix Z is the impedance operator Z represented in {ψ n } basis.”

From Impedance Operator Z to Impedance Matrix Z

The impedance matrix Z reads

I We say8 : “Matrix Z is the impedance operator Z represented in {ψ n } basis.”

Two Hilbert Space Representations12

Example: Complex Power Balance

I Notice Prad,n > 0 ⇒ R 0 is required.